JPH086948B2 - Control device for air conditioner - Google Patents

Description

Description: TECHNICAL FIELD The present invention is to automatically perform comfortable air conditioning operation in an air conditioner equipped with a microcomputer, for example, to control indoor temperature, air volume and wind direction. The present invention relates to a control device of an air conditioner for improving the comfort of a person in a room.

2. Description of the Related Art When controlling the room temperature with an air conditioner, in the heating example, control is performed to shift the indoor target temperature of the air conditioner to a higher value for a certain period of time in order to improve the characteristics when the room temperature rises. The method of controlling the machine operating frequency was adopted. FIG. 6 shows such a conventional air conditioner control device, in which the control signal generation means 61 is a suction temperature 65 from the sensor 60 and a timer from a timer 63 which operates after the air conditioner 62 is powered on. Value 67 and remote control or operation panel 64 for operating the air conditioner 62 from the outside
The control signal 66 is generated by the user-set temperature 68 or the like. For example, during heating, within 60 minutes after the power is turned on, the indoor target temperature is set to 2 ° C higher than the user's set temperature set by the remote control or operation panel 64 in order to quickly raise the indoor temperature. The room temperature adjustment of the air conditioner is controlled so as to be set.

DISCLOSURE OF THE INVENTION Problems to be Solved by the Invention However, in the above-described conventional control device, since the control is performed only by the time after the power is turned on and the indoor temperature characteristic, the magnitude of the air conditioning load in the room in which the air conditioner is installed is small or large. Can't deal flexibly with. Therefore, for example, when the load is too low, the room temperature becomes higher than the target temperature, and when the load is too high, it takes a considerable time for the room temperature to reach the target temperature, and it is not possible to consider the comfort of human beings in the room. There is a problem.

In view of the conventional problems as described above, the purpose of the present invention is to
(EN) Provided is a control device for an air conditioner, which can realize a more comfortable air conditioning and a living environment in consideration of the comfort of a person in a room.

Means for Solving the Problem In order to achieve this object, the present invention provides a plurality of sensor means for detecting indoor and outdoor temperature, air volume of an air conditioner, and humidity as indoor and outdoor environmental conditions, As a state, human comfort is obtained from storage means for storing the intake air temperature gradient of the air conditioner at intervals of N seconds (N is a positive real value), the output from the sensor means and the storage means, and the temperature set by the user. Using the neural network model means (neural network) that has learned the feeling, the human metabolic rate as the indoor comfort feeling, which is an evaluation index for controlling the air conditioner,
Predicted average number of votes (PMV) calculated by at least one of human clothing condition, room temperature, humidity, air flow velocity, or ambient wall radiant temperature, or standard new valid for predicting human physiological condition and sensation Estimating means for estimating temperature (SET),
Control of an air conditioner, comprising: a control signal generation unit that generates a control signal such as a blowing temperature, a wind direction, and an air volume of the air conditioner based on a comfort feeling of a person in the room estimated by the estimating unit. It proposes a device.

Operation According to the above-mentioned configuration of the present invention, the estimating means determines whether the indoor condition is determined by the indoor and outdoor environmental conditions detected by the plurality of sensor means, the previous state of the sensor means held in the storage means, and the temperature set by the user. Guess the human comfort level. A control signal is generated by the control signal generation means on the basis of the comfort of the human being estimated by the estimation means to control the air conditioner. With this, considering the indoor environment and human condition,
A more comfortable air conditioning environment can be realized.

Example Hereinafter, an example of the present invention will be described in detail with reference to FIGS. 1 to 5.

FIG. 1 is a block diagram showing the flow of signals in the control device according to the first embodiment of the present invention, and FIG. 2 is a block diagram showing the learning method of the neural network schematic means in FIG. In FIG. 1, 10 is a sensor, 11 and 12 are sensor signal values from the sensor 10, 13 is a storage means, 14 is a slope of the suction temperature output from the storage means 13 at N second intervals, 15
Is a remote control or operation panel, 16 is an output signal from the remote control or operation panel 15, 17 is a neural network schematic means, 18
Is the comfort level (Predicted Mean Vote, hereinafter referred to as PMV) or Standard Effective Temperature (hereinafter referred to as SET) output from the means 17.
, 19 is a control signal generating means, 1a is a control signal output from the means 19, and 1b is an air conditioner.

Next, the learning method of the neural network schematic means will be described with reference to FIG. 2. In FIG. 2, 21 is the outdoor temperature, 22 is the suction temperature, 23 is the slope of the suction temperature, 24 is the air volume, and 25 is the user's air temperature. Set temperature, 26 is human temperature, 27 is neural network model means, 28 is control signal generation means, 29 is actually measured PMV (or SET), 2a is estimated PMV (or SET), 2c is control signal, 211, 221, 231, 241, 251, 261 is Outdoor temperature 21, respectively
The signals are the suction temperature 22, the suction temperature gradient 23, the air volume 24, the user set temperature 25, and the human body temperature 26.

Since the first embodiment of the present invention is configured as described above, the sensor signal 11 is output from the plurality of sensors (outside air temperature sensor, suction temperature sensor, humidity sensor, human body temperature sensor) 10 in the air conditioner 1b. Will be done. This signal 11
Is the outdoor temperature, the suction temperature, the humidity, the human body temperature, and the like.

Further, a signal 12 similar to the signal 11 is output from the sensor 10 and input to the storage means 13, but the storage means 13 receives the sensor output signal 12 for the past N seconds (N is a positive real number). Remember history. Remote control or operation panel
Since the air volume and the set temperature value 16 of the user are output from 15,
Further, the storage means 13 outputs the slope 14 of the room temperature at intervals of N seconds (N is a positive real number) from the state before the sensor, for example, the history of N seconds described above. Output signal from each means 10, 13, 15
11, 14, 16 will be input to the neural network model means 17. This neural network model means 17 outputs the predicted average number of votes PMV, which is the comfort level in the room, or the estimated value 18 of the standard new effective temperature SET from the input signal. PMV is an environmental test room in which the combination of 6 elements such as temperature, humidity, air flow velocity, radiant temperature (surrounding wall), metabolic rate, and clothing state is changed as factors that affect comfort. It was quantified based on the results of the polls about the cold and hot weather. In other words, the calculated PMV values are −3: cold −2: cool −1: a little, depending on the human condition (metabolic and clothing conditions) and the indoor environment (temperature, humidity, air flow velocity, ambient wall radiation). Cool 0: Nothing +1: Somewhat warm +2: Warm +3: Can be evaluated as hot.

On the other hand, SET calculates the amount of thermal stimulation from physical factors of the environment,
It seeks to predict human physiological state values and sensations,
It is one of the physical evaluation methods for comfort that captures the relationship between pleasantness and discomfort with heat by the physiological response to the physical quantity of thermal stimulation. For example, in the case of using PMV, by inputting the human condition such as the outdoor temperature, the suction temperature, the suction temperature gradient, the air volume, the user's set temperature, and the human body temperature into the neural network model 17, the neural circuit is input. Net model means 17 to PM
The estimated value of 18 is output. The estimated value 18 of the PMV is input to the control signal generating means 19, and the control signal is generated from the generating means 19.
1a is generated. The control signal generation means 19 will generate the control signal 1a that maximizes the performance of the air conditioner 1b when the comfort level is unsatisfactory. In this case, if the comfortable feeling is further satisfied, the control signal 1a is generated so that the comfortable feeling can be maintained. That is, although the signal 1a for controlling the air conditioner 1b is generated by the estimated PMV value 18, the control signal 1a controls the inverter frequency, the wind direction, the air volume, the indoor target set temperature, etc. in the air conditioner 1b. It

As an example, the shift amount for calculating the indoor target temperature targeted by the air conditioner 1b is obtained from the outside air temperature, the intake temperature, the air volume, the set temperature, the inclination of the intake temperature, etc. from each of the means 10, 13, and 15. However, the relationship between this shift amount and the temperature set by the user and the indoor target temperature is: indoor target temperature = user set temperature + shift amount. Therefore, the control signal 1a of the air conditioner 1b generated by the control signal generating means 19 is input to the air conditioner 1b, and the operation of the air conditioner 1b is executed so as to reach the indoor target temperature based on the above (formula), for example. It

 Thereby, the indoor temperature adjustment 1c is performed.

Next, details of the learning method of the neural network of FIG. 1 will be described with reference to FIG. In FIG. 2, signals from the outdoor temperature 21, the intake temperature 22, the inclination 23 of the intake temperature at N second intervals, the air volume 24, the set temperature 25 of the user, the human body temperature 26, etc.
When 211, 221, 231, 241, 251, and 261 are input to the neural network schematic unit 27, the estimated value 2a of PMV is output.

The neural network pattern model means 27 learns the measured value 2a of PMV by using the actually measured PMV 29 measured indoors as the learning data 2b. When the actually measured PMV 29 uses the amount of clothes and the amount of metabolism of the user as the human state, the amount of clothes and the amount of metabolism obtained from the user's set temperature are used. That is, if the set temperature is low, it can be inferred that the amount of clothes is large and the amount of metabolism is high, and conversely, if the set temperature is high, it can be inferred that the amount of clothes is small and the amount of metabolism is low. If there is a human body temperature sensor, a value obtained by similarly estimating the amount of clothing and the amount of metabolism from the information of the human body temperature sensor is used.

There are various methods of learning neural network learning algorithms, for example, backpropagation algorithm (reference: Ramelhart, DE and McClellan. JL "PDP model-cognitive science and search of neuron network") {Runmelha
rt, DE and Mcclelland, JL (Eds.), Parallel Distri
buted Processing, Exploration in the Microstructure
of Cognition.Vol.1,2, MIT Press, Cammbridge (198
6)}) is used to obtain the optimal solution by the method of maximum descent. Then, the learning is performed by these algorithms until PMV can be sufficiently estimated by the neural network schematic means 27. When learning is completed, the output value 2a of the neural network schematic means 27 causes
When the comfort is not satisfied, the control signal generating means 28 generates the control signal 2c that maximizes the capacity of the air conditioner. Further, when the comfortable feeling is satisfied, the control signal 2c is generated by the control signal generating means 28 so that the comfortable feeling can be maintained. That is, the signal 2 for controlling the air conditioner based on the PMV value 2a estimated by the neural network schematic means 27
c is generated. The control signal 2c controls the inverter frequency, the wind direction, the air volume, the indoor target set temperature, and the like.

As can be understood from the above description, according to the present embodiment, the input from each sensor is input to the neural network schematic means, the PMV is estimated, and the control signal is generated by the value of the PMV to generate the indoor environment. In consideration of the above, a more comfortable air conditioning environment can be realized.

In this embodiment, the outputs 18 and 2a from the neural network schematic units 17 and 27 are PMV, but the same effect can be obtained by rewriting to SET or ambient wall radiation temperature.

Next, a control device for an air conditioner according to a second embodiment of the present invention will be described with reference to FIGS. 3 to 5.

FIG. 3 is a block diagram showing a signal flow in the control device according to the second embodiment of the present invention, in which 30 is a sensor,
31, 32 are sensor signal values from the sensor 30, 33 is storage means, 3
Reference numeral 4 is a slope of the suction temperature output from the storage means 33 at intervals of N seconds, 35 is a remote control or operation panel, 36 is an output signal from the remote control or operation panel 35, 37 is a look-up table, and 38 is a look-up table 37. The output control signal 39 indicates the air conditioner.

Next, the lookup table creation method will be described with reference to FIG. 4. 41 is the outdoor temperature, 42 is the intake temperature, 43 is the slope of the intake temperature, 44 is the air volume, 45 is the user's preset temperature, 46 is the human body temperature, 47 is a neural network schematic means, 48 is a control signal generating means, 49 is a measured PMV (or SET), 4a is a PMV (or SET) estimated from the neural network schematic means, 4b is a measured PMV (or SET) Comfort level (PMV or SET),
4c is a control signal, 4d is a lookup table, 411,421,43
1,441,451,461 are outdoor temperature 41, suction temperature 4 respectively
2, gradient of suction temperature 43, air volume 44, user set temperature 4
5 shows the signals from the human body temperature 46.

Since the air conditioner according to the second embodiment is configured as described above, the sensor signals 31 are output from the plurality of sensors 30 in the air conditioner 39. This signal 31 is the outside temperature,
Inhalation temperature, humidity, human body temperature, etc.

A signal 32 similar to the signal 31 is output from the sensor 30, and this signal is input to the storage means 33. Storage means 33
Stores the history of the input sensor output signal 32 for the past N seconds (N is a positive real number). The air volume and the set temperature value 36 of the user are output from the remote control or the operation panel 35, and the storage means 33 outputs the state before the sensor, for example, the slope 34 of the room temperature at intervals of N seconds (N is a positive real number). . The output signals 31, 34, 36 from the respective means 30, 33, 35 are input to the look-up table 37 as input signals, and the look-up table 37 determines the control signal 38 for the air conditioner 39. The control signal 38 controls the inverter frequency, the wind direction, the air volume, the indoor target set temperature, and the like in the air conditioner 39. As an example, the shift amount for calculating the indoor target temperature targeted by the air conditioner 39 is obtained from the outside air temperature from each of the means 30, 33, and 35, the intake temperature, the air volume, the set temperature, the inclination of the intake temperature, and the like. To be The relationship between the shift amount and the temperature set by the user and the indoor target temperature is: indoor target temperature = user set temperature + shift amount. Therefore, the control signal 38 obtained from the look-up table 37 is input to the air conditioner 39, and the operation of the air conditioner 39 is executed so that the indoor target temperature is reached, for example, based on the above formula.

Next, referring to FIG. 4, the method of creating the lookup table of FIG. 3 will be described. The outdoor temperature 41 and the suction temperature 42, the suction temperature gradient 43 at N second intervals, the air volume 44, the user set temperature 45, and the human body temperature. Signals from 46 mag 411,421,43
Input 1,441,451,461 to neural network model 47,
The estimated value 4a of is output.

The neural network model means 47 learns the estimated value 4a of PMV using the actually measured PMV 49 measured indoors as the learning data 4b.

There are various methods for learning algorithms of the neural network, and for example, the optimal solution is obtained by the descent method by the above-described backpropagation algorithm. With these algorithms, PMV is sufficient for neural network schematic means 47
Learning is performed until it can be estimated with, but when the learning is completed, the output value 4a of the neural network schematic unit 47
When the comfort is not satisfied, the control signal generation means 48 generates the control signal 4c that maximizes the capacity of the air conditioner. Further, when the comfortable feeling is satisfied, the control signal 4c is generated by the control signal generating means 48 so that the comfortable feeling is maintained.

The control signal 4c controls the inverter frequency, the wind direction, the air volume, the indoor target set temperature, etc., but the look-up table 4d includes the neural network schematic means 47 and the control signal generating means 48 for outputting the control signal 4c. In order to replace the input signals 411 to 461 of the outdoor temperature 41 to the human body temperature 46, which are sensor inputs, with a rough quantization, the lookup table 4d
To the look-up table 4d to create a look-up table. Fifth
The figure shows one embodiment of the lookup table 4d, which includes zones A to C, in which the set temperatures ti to te, the outside air temperature toi to tom, the air volume fi to fn, and the suction temperature si to. so and the suction temperature gradient ki to kp will be written.

As described above, according to the present embodiment, the input from each sensor is input to the neural network schematic means, the PMV is estimated, and the control signal is generated according to the PMV value to consider the indoor environment. It is possible to realize a more comfortable air conditioning and living environment. Furthermore, from the neural network model means
The control device can be easily realized by estimating the PMV and replacing the part for converting into the control signal with the look-up table.

EFFECTS OF THE INVENTION As described above, according to the control device for an air conditioner of the present invention, a more comfortable air conditioning environment can be realized in consideration of the indoor environment and the human condition. Cheaper,
Obviously, it can be easily implemented.

[Brief description of drawings]

FIG. 1 is a block diagram showing a flow of signals of a control device for an air conditioner according to a first embodiment of the present invention, and FIG. 2 is a block diagram for explaining a learning method of neural network schematic means in FIG. 1, FIG. 3 is a block diagram showing a signal flow of a control device for an air conditioner according to a second embodiment of the present invention, and FIG. 4 is a block diagram for explaining a method of creating the lookup table in FIG. FIG. 5 is a diagram showing an example of the lookup table in FIGS. 3 and 4, and FIG. 6 is a block diagram showing a conventional control device. 10,30 …… Sensor, 11,12,31,32 …… Sensor signal, 13,33
...... Memory means, 14,34 ...... Suction temperature gradient every N seconds, 15,35 ...... Remote control or operation panel, 16,36 ...... Air volume, set temperature, etc. 17 ...... Neural network model, 18 ...... Comfort estimated value, 19 ... Control signal generation means, 1a, 38 ... Control signal, 1b, 39 ... Air conditioner, 1c, 3a ... Indoor temperature adjustment, 2
1,41 …… Outdoor temperature, 22,42 …… Suction temperature, 23,43 ……
Gradient of suction temperature, 24,44 …… Air volume, 25,45 …… Set temperature of user, 26,46 …… Human body temperature, 27,47 …… Neural network model means, 28,48 …… Control signal generation Means, 29,49 …… Measurement
PMV, 2a, 4a …… PMV estimated value, 2b, 4b …… PMV learning data, 2
c, 4c …… control signal, 37,4d …… lookup table, 6
0 …… Sensor, 61 …… Control signal generating means, 62 …… Air conditioner, 63 …… Timer, 64 …… Remote control or operation panel, 65 …… Suction temperature, 66 …… Control signal, 67 …… Timer value , 68 …… Set temperature.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Takehisa Tanaka 3-10-1 Higashisanda, Tama-ku, Kawasaki City, Kanagawa Prefecture Matsushita Giken Co., Ltd. 3-10-10 Mita Matsushita Giken Co., Ltd. (56) Reference JP-A-4-43248 (JP, A) JP-A-3-110340 (JP, A)

Claims (2)

[Claims] 1. A plurality of sensor means for detecting indoor and outdoor temperatures, air volume of an air conditioner, and humidity as indoor and outdoor environmental conditions,
N seconds as the previous state of the sensor means (N is a positive real value)
First storage means for storing the intake air temperature gradient of the air conditioner at intervals, neural network schematic means for learning human comfort from the outputs from the sensor means and the storage means and the temperature set by the user (Neural network) is used as an indoor human comfort level, which is an evaluation index for controlling the air conditioner, as human metabolic rate, human clothing, indoor temperature, humidity, air flow velocity, or ambient wall radiation. Prediction mean votes (PMV) calculated by at least one of the temperatures, or guessing means to guess the standard new effective temperature (SET) that predicts human physiological condition and sensation, and the guessing from the guessing means An air conditioner control device, comprising: a control signal generating unit that generates a control signal such as a blowing temperature, a wind direction, and an air volume of the air conditioner based on the comfort of a person in the room. 2. The estimating means and the control signal generating means are a second storing means comprising a sensor means and a look-up table for generating a control signal from the output from the storing means and the temperature set by the user. The control device for the air conditioner according to claim 1.